2nd International Conference on Advances in Experimental Structural Engineering
Tongji Univ., Shanghai 200092, China
December 4-6, 2007
EXPERIMENTAL TESTS ON MASONRY STRUCTURES: FROM THE
SIMPLEST IN-SITU TEST TO COMPLEX LAB TEST SETUPS
Alexandre A. Costa1, António Arêde 2 and Aníbal Costa 3
1
Faculty of Engineering of the University of Porto, Portugal, aacosta@fe.up.pt
Faculty of Engineering of the University of Porto, Portugal, aarede@fe.up.pt
3
Departamento de Engenharia Civil, University of Aveiro, Portugal, agc@ua.pt
2
Abstract
This paper presents different experimental test techniques used to assess the behavior of different
masonry structures. The test setups presented can be considered from the simplest one for in-situ use,
to more complex ones used in laboratory conditions. Some examples are presented as used to assess
the in-plane and out-of-plane performance of masonry walls subjected to loads using simple in-situ or
complex laboratory setups. In particular, an experimental test for behavior assessment of a masonry
arch bridge is referred as an example of a simple test methodology with complex instrumentation.
Finally, some comments are included regarding the efficiency of simple test schemes when compared
to more complex and elaborated test setups as well as the applicability of the later to in-situ or
laboratory conditions regarding costs, time and the need for their use.
Keywords: Masonry walls, Arch bridge, Experimental, In-situ tests, Laboratory tests, Monitoring
1. Introduction
Experimental tests are a major issue on structural engineering due to the relevant outcome for design
and assessment purposes. Several different test types can be performed on elements/structures, such as
material characterization, fatigue, cyclic behavior or even seismic performance assessment of
complete structures.
However requirements to perform complex tests are not always available mainly due to technical
issues and costs associated with the involved procedures. Hence, engineered solutions can be arranged
in order to obtain equivalent experimental results using alternatives to the desired test setups.
In this context, different test procedures to assess the behavior of elements and structures will be
briefly presented based on experimental activities already carried out with different load conditions,
control procedures and monitoring techniques; more relevance is given to the most innovative
solutions.
2. Experimental Tests on Masonry Panels
Testing of masonry panels is a key issue to allow predicting their performance under seismic
excitations, particularly because under simple quasi-static loading it is possible to estimate the cyclic
response envelope. Actually, this can be done by recourse to experimental tests on masonry panels
subjected to vertical forces (simulating the mass of a structure above it) and a horizontal load applied
at a given height in order to simulate the inertial excitation due to an earthquake action. Different ways
- 485 -
can be adopted to apply these loads, namely monotonic, cyclic and earthquake simulation as detailed
in Tomazevic [1] and Costa [2].
Additionally, reproducing adequate restrain conditions is another issue playing a major role in this
type of experimental tests. While it is possible to simulate a cantilever response of a masonry pier or a
fixed-ended condition on a masonry panel, as mentioned by Costa [2], it is also clear that in-situ
boundary conditions are very hard to reproduce in laboratory and, therefore, the best way definitely
consists on playing with in-situ testing.
2.1. In-plane Behaviour
In order to assess the lateral capacity of masonry piers with a simple test setup, a shear load should be
applied on the specimen. However this shear load can be applied in several different ways either in the
laboratory or in-situ.
The first case presented herein refers to a test setup recently used at the University of Aveiro
(Portugal), consisting of a very simple system shown in Fig. 1 a), in which a simple load-handling
mechanism attached to a steel cable was used to impose lateral forces by cyclic varying the cable
tension but without reversals. The horizontal force was monitored using a digital dynamometer,
whereas the vertical force was applied at the top resorting to dead load using controlled weight cement
bags. Displacements were controlled using mechanical dial gauges rather than electronic devices.
This test setup, in what concerns horizontal loading, was inspired on a process used for the profile
correction recently carried out on a large masonry wall of the Tibães Monastery
(http://www.mosteirodetibaes.org/) near Braga, Portugal, that was monitored by the Laboratory of
Earthquake and Structural Engineering (LESE) of the Faculty of Engineering of University of Porto
(FEUP). The same loading system was adopted there, though with more powerful manual devices
(tirefort) and the idea was subsequently used in the assessment of the in/out-of-plane capacity of a
masonry arch pier in the Aveiro Museum (Fig. 1 b)), taking advantage of a stiffer wall to act as
“reaction wall” (see Fig. 1 c)).
From Fig. 1 it is clear that this system is a low cost and very simple test setup that can be used either in
the lab or in-situ, adopting any kind of monitoring devices.
(a)
(b)
(c)
Figure 1: Simple setup for in-plane testing of masonry panels: (a) at lab facilities, (b) and (c) in-situ.
Typical experimental laboratory tests on masonry panels can be considered at an intermediate level of
complexity because simplified test setup apparatus and simple PID controllers are often used. Results
obtained from such tests are very helpful on studying carefully masonry elements and for developing
numerical behavior models, mainly because the specimens can be fully instrumented and precisely
displacement-controlled using hydraulic actuators in laboratory controlled conditions. An example of
such a test setup is shown in Fig. 2 referring to the system currently used at LESE-FEUP for testing
masonry walls in a typical cantilever test setup (as explained by Costa [2]).
- 486 -
Another case, referring to the experimental test setup currently in use at the EUCENTRE facilities
(Pavia, Italy) is presented in Fig. 3. The system was designed to assess the in-plane behavior of
masonry panels under double bending condition [2] and it can be considered a complex test setup
because it makes use of a mixed force/displacement controller on the two vertical actuators to apply
the vertical load and to ensure the horizontality of the top beam. The test may be considered stable and
valid if the conditions (1) and (2) expressed in Fig. 3 a) are met, where NT is the total axial force and
(NA , δA) and (NB , δB) refer to the forces and displacements on actuators A and B, respectively. Despite
the complexity of the test setup controller, this is a very simple apparatus allowing any kind of
measurement devices, fast operations between experiments, clear view of the specimen and, last but
not the least, safe even with the collapse of the specimen [2].
Figure 2: Test setup currently in use at FEUP (Faculty of Engineering of the University of Porto).
NT = N A + N B (1)
δA = δB
(2)
(a)
(b)
Figure 3: Complex test setup at EUCENTRE, Pavia, Italy: (a) scheme, (b) testing a masonry panel [2].
2.2. Out-of-plane Behaviour
The out-of-plane motions of masonry walls during earthquake are commonly associated with partial or
fully collapse of masonry panels or even buildings, as recently observed in Bam 2003 or Pakistan
2005 earthquakes. Indeed, experimental tests general related to this topic are essentially based on
shaking table tests, airbag tests (Griffith [3]) and waterbag tests (Mosallam [4]), aiming at simulating
the panel distributed load due to seismic excitation. However, some experimental tests with
concentrated loads on the masonry panel have been also performed, as reported in Milani et al. [5].
Despite the fact that all these experimental tests were successfully done and implemented in laboratory
conditions, they are not easy to be done in-situ due to the test nature and because reaction structures or
shaking tables are required.
Bearing this in mind, and also the need for gathering realistic information of the behavior of structures
as they exist, two different test setups were already used in-situ by the authors in order to simulate the
seismic action on masonry panels of existing constructions, as described below.
- 487 -
The first setup presented in this paragraph refers to in-situ tests carried out in the Faial island, Azores,
Portugal, after the Azores 1998 earthquake, during an experimental test campaign by Costa [6]. The
test setup consisted on two steel buckets that could be filled with sand and able to vibrate freely,
suspended by a steel cable to an auxiliary steel bar structure attached to the wall specimen (Fig. 4).
(a)
(b)
(c)
(d)
Figure 4: In-situ test setup used by Costa [6]: (a) site 1, (b) site 2, (c) site 3, (d) monitoring devices.
With this system, horizontal out-of-plane forces were imposed to the wall by applying vertical loads in
the buckets. In addition, load cycling was also made possible by means of forced oscillation of the
bucket and by playing with the phase angle of two bucket cyclic motion, one in each side of the wall.
Fig. 5 thus presents the force plot from the two steel buckets oscillating with a small phase angle.
200
Force 1
Force 2
Imposed load
150
Force (kN)
100
50
0
0
5
10
15
20
25
30
-50
-100
-150
-200
Time step
Figure 5: Possible imposed loads by Costa [6] test setup.
Associated with this simple test setup, also simple monitoring devices were used for quick and easy
installation. Despite the adoption of simple mechanical dial gauges to obtain the out-of-plane
displacements as shown in Fig. 4 d), it is apparent that any other type of other transducers could have
been used in this test setup. However, it is also clear that this test setup is very limited regarding the
load itself because it is “force controlled” with a limited load pattern.
Recently, an experimental test campaign was performed in Azores with more sophisticated test setup
and monitoring devices. Indeed a portable and powerful test scheme was developed (as stated by
Arede et al. [7] and Costa et al. [8]) where hydraulic jacks were used to apply the desired loads and
draw wire position transducers (DWPT) were selected to monitor out-of-plane movements, with a
range of 150-625mm and a linearity of ±0.1% of full scale (WayCon™). These instruments were
controlled and monitored using a laptop and two small USB devices (National Instruments™ USB6210 and USB-6211 with a maximum sample rate of 250 ksamples/sec), which allows to handle a
total of 16 differential input channels (or 32 non-reference single ended input channels), two
differential output channels and 5 digital channels. Finally the applied force was monitored using a
load cell on the tip of one hydraulic jack.
All these instruments (data acquisition and control devices) were controlled by a computer code
developed at LESE-FEUP allowing real-time control and acquisition based on the LabVIEW software.
Fig. 6 presents all the equipment used to perform such tests.
- 488 -
(a)
(b)
(c)
Figure 6: Equipment used by Arêde et al. [7]: (a) acquisition and control devices, (b) hydraulic power unit
and actuators, (c) DWPTs.
In this new experimental test setup two parallel walls were tested simultaneously, based on an actionreaction scheme. Indeed the hydraulic jacks were placed in between and connected to these walls at
the roof/floor level in order to simulate the floor/roof effect and the masonry inertial forces, as it is
represented in Fig. 7.
0.36
0.82
0.15
0.66
0.15
1.00
2.01
0.65
0.85
1.93
2.01
0.62
1.45
1.41
0.94
1.04
LOCAL 2
1.18
0.65
1.18
2.93
1.36
1.04
0.95
1.21
1.18
1.09
0.15
0.99
3.43
1.68
11.82
1.11
LOCAL 1
1.71
1.03
3.22
1.35
2.01
(a)
(b)
(c)
(d)
Figure 7: Experimental test setup: (a) reproduced effect, (b) in-situ plan view, (c) local 1 (d) local 2.
Using this test setup it is possible to study simultaneously the out-of-plane behaviour of two walls in
the same experimental test, monitoring both walls at the same time. This means that it is possible to go
up to the collapse of the weakest wall in this in-situ experimental test setup because the stiffer one acts
as “reaction wall” up to the end of the test.
3. Masonry Bridge Case
The structural behaviour of stone masonry bridges is still a challenging and pertinent topic of research,
although it typically refers to old structures. The very fact that a large number of such constructions
exist, in Portugal as in many other countries, under service conditions for which they were not
designed, increases the problem relevance [9].
The following case study refers to the new stone masonry bridge in Vila Fria - Portugal, over the
Vizela River, recently built according to traditional techniques of masonry construction to replace an
old and very deficient passage (Fig. 8).
The bridge is 6 m wide and has a total span of 60 m, distributed by five stone arches of 4.8 m to 6.0 m
long and two abutments. The main structural elements, namely arches, piers and spandrel walls are
fully made of stone masonry brickwork, whereas the bridge interior is made of filling material
composed by a selected tout-venant.
3.1 Instrumentation and Data Acquisition
- 489 -
By taking profit of the construction process a large instrumentation network was installed in the bridge,
including both electrical type and fibre optic based sensors, the later particularly relying on fibre
Bragg gratings [10].
(b)
(a)
Figure 8: The Vila Fria bridge: (a) old passage and (b) new stone masonry bridge.
Strain in stone blocks is measured in 48 points by electrical and optical strain gauges (Fig. 9); since
this type of structure has very low stress level in the linear elastic domain, the strain measures give
directly an indication of the stresses installed.
Section view 1
E6I
(a)
(b)
(c)
Figure 9: Strain gauge layout: (a) downstream façade alignment (similar for the central alignment), (b)
electrical strain gauge in the respective stone slot and (c) gauge location in the extrados.
Pressure distribution in a selected region is monitored by 7 total electrical pressure cells [11] as
usually adopted in geotechnical works (Fig. 10). Vertical movements are obtained using 16 ultra-low
differential electrical pressure sensors from Honeywell with an appropriate silicone oil to measure
differential displacements between a series of points along the bridge deck. Relative opening/closing
and slip displacements in joints between blocks, as well as relative displacements between spandrel
walls are monitored by means of 53 displacement transducers based on fibre optic sensors; the same
technology has been adopted for temperature measurement (28 sensors).
(a)
(b)
(c)
Figure 10: Total pressure cells: (a) location, (b) circular cell (pier-infill) and (c) half-moon cell (inside
stone blocks).
- 490 -
Data of all the sensors are collected by appropriate acquisition systems, namely one BraggMeter
measurement unit containing an optical switch from FiberSensing and one Compact Fieldpoint
acquisition unit from National Instruments. Both equipments are connected by TCP/IP to a router
allowing wireless data transmission by GPRS/UTMS to a remote server at FEUP where monitoring
results are collected in a specifically designed database and accessible through an internet site
(http://remotelese.fe.up.pt/). The whole monitoring system is described in detail in [12] and [13].
3.2 Load Test
After completion of all the activities related to instrumentation, calibration and whole system testing,
the load test on the bridge was carried out using four trucks loaded to their maximum capacity around
39 ton (40% for the two front axles and 60% for the two rear axles). Three vehicle arrangements were
adopted to obtain the most unfavourable structural effects [13]. In order to obtain a more concentrated
load, the arrangement 1 consisted in just one pair of side-by-side trucks in twelve stopping positions
along the whole bridge length defined by the middle position of the two rear axles. Fig. 11 shows the
stopping positions for the trucks along the bridge for the three arrangements (1A to 1M, for the first;
2A to 2F for the second; 3A to 3C for the third), as well as a photo of one test stage (3A, with trucks
all over the full length of bridge arches).
(a)
(b)
Figure 11: Load test: (a) truck stopping positions and (b) truck arrangement for position 3A.
The arrangement 1 was found to produce the most unfavourable effects. The most expressive results
refer to the pressure values obtained from pressure cell readings that can be seen in Fig. 12: note the
most unfavourable situation for position 1F right above the arch crown, as expected. Vertical
alignments of these plots correspond to the several stopping positions and are identified in Fig. 12.
Positive values refer to increase of compression that are detected in the arch 4, the most instrumented
one and likely to exhibit more pronounced effects. A general overview of the observed behaviour
confirms the large stiffness of the bridge that is consistent with the very low deformations recorded:
0.31mm for the vertical deflection at the crown of arch 4 and joint closure of 0.16mm in the same zone.
200
Início
1B
1D
1E
1F
1G
1H
1I
1J
1K
1L
1M
Fim
150
100
50
0
16:25:05
16:39:29
16:53:53
17:08:17
17:22:41
17:37:05
Figure 12: Load test results for arch 4 and load arrangement 1 - Stress evolution (kPa) in load cell C7, for
all positions of the first vehicle arrangement.
- 491 -
Long term monitoring results are already available since May 2006, although not yet processed and
analysed; they are expected to provide information regarding the influence of temperature on the
sensors’ response.
At the same time several numerical simulations were performed to simulate the response of the load
test, using the general purpose computer code CAST3M [14] and finite element modelling strategy as
reported in [12] and [15].
4. Conclusions
The significance of in-situ tests was remarked in this work due to their own characteristics and
outcome. They are really important to study elements/structures under their natural conditions, as
vertical loads, structural systems, materials, connections between materials and surrounding structures,
etc. Despite the great relevance of laboratory experimental tests, where specimen testing is done in
controlled environments as required to develop and calibrate numerical models, in-situ tests also play
a major role in order to use such calibrated models and to adjust them to real conditions, thus finally
allowing assessing the behavior of existing structures. Indeed, as shown below, it is possible to carry
out low cost experimental tests with significant data outcome to study elements/structures’ behavior.
Moreover the continuous monitoring of new/existing structures is increasing due to the invaluable data
they can provide over time, quite relevant to understand the evolution of structures’ performance and
to allow repairing/strengthening/retrofitting on due time.
References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
M. Tomazevic, Earthquake-resistant design of masonry buildings, Imperial College Press, 1999.
A.A. Costa, Experimental Testing of Lateral Capacity of Masonry Piers, An Application to Seismic
Assessment of AAC Masonry Buildings, MSc Thesis on Earthquake Engineering, European School for
Advanced Studies in Reduction of Seismic Risk (ROSE School), Pavia, Italy, 2007.
M.C. Griffith, J. Vaculik, N.T.K. Lam, J. Wilson and E. Lumantarna, “Cyclic testing of unreinforced
masonry walls in two-way bending,” Journal of Earthquake Engineering and Structural Dynamics, Vol.36,
No.6, 801-821, 2007.
A.S. Mosallam, “Out-of-plane flexural behavior of unreinforced red brick walls strengthened with FRP
composites,” International Journal of Composites Part B:Engineering, Vol.38, 559-574, 2007.
G. Milani, F.A. Zuccarello, R.S. Olivito and A. Tralli, “Heterogeneous upper-bound finite element limit
analysis of masonry walls out-of-plane loaded,” Journal of Computational Mechanics, Vol.40, No.6, 911931, 2007.
A. Costa, “Determination of mechanical properties of traditional masonry walls in dwellings of Faial Island,
Azores,” Journal of Earthquake Engineering & Structural Dynamics. Vol.31, No.7, 1361 – 1382, 2002.
A. Arêde, A. Costa, A.A. Costa, C.S. Oliveira and F. Neves, “Caracterização Experimental do Sistema
Construtivo do Grupo Central do Arquipélago Açoreano – Uma Contribuição para a Avaliação e Redução
da sua Vulnerabilidade Sísmica,” SÍSMICA 2007 – 7º Congresso de Sismologia e Engenharia Sísmica,
FEUP, Porto, 2007 (in Portuguese)
A.A. Costa, A. Costa, A. Arêde, C.S. Oliveira and F. Neves, “Metodologia da preparação de ensaios a
realizar na ilha do Faial, Açores,” SÍSMICA 2007 – 7º Congresso de Sismologia e Engenharia Sísmica,
FEUP, Porto, 2007 (in Portuguese).
C. Costa, Análise do Comportamento da Ponte da Lagoncinha sob a Acção do Tráfego Rodoviário. MSc
Thesis on Civil Engineering, Faculdade de Engenharia da Universidade do Porto (FEUP), 2002 (in
Portuguese).
L.A. Ferreira, F.M. Araújo, A.A. Maia, P.A. Alves and J.L. Santos, “Redes de Bragg em Fibra Óptica Contribuições para o desenvolvimento de uma tecnologia revolucionária em monitorização estrutural”,
Betão Estrutural 2004, 1031-1038, FEUP, Porto, 2004.
Geokon, Inc., Earth Pressure Cells , 2002.
A. Arêde, P. Costa, A. Costa, C. Costa and L. Noites, “Monitoring and testing of a new stone masonry arch
bridge in Vila Fria, Portugal,” Proceedings of Arch’07, Funchal, 2007.
P. Costa, Análise da Construção e do Comportamento de uma Ponte de Pedra, MSc thesis on Civil
Engineeering, Faculdade de Engenharia da Universidade do Porto (FEUP), 2007 (in Portuguese).
P. Pasquet, “Manuel d'utilisation de Cast3m,” www-cast3m.cea.fr. Commissariat à l'Énergie Atomique,
2003.
C. Costa, P. Costa, A. Arêde and A. Costa, “Structural design, modelling, material testing and construction
of a new stone masonry arch bridge in Vila Fria, Portugal,” Proceedings of Arch’07, Funchal, 2007.
- 492 -